(* Copyright NICTA 2008 *)
(* Distributed under the terms of the LGPL (see LICENCE) *)
(******************************************************************************)
-(*INCLUDE "debug.ml" *)
+INCLUDE "utils.ml"
type tree
type 'a node = int
external parse_xml_string : string -> int -> bool -> bool -> tree = "caml_call_shredder_string"
external save_tree : tree -> string -> unit = "caml_xml_tree_save"
-external load_tree : string -> int -> tree = "caml_xml_tree_load"
+external load_tree : string -> int -> tree = "caml_xml_tree_load"
external nullt : unit -> 'a node = "caml_xml_tree_nullt"
let nil : 'a node = Obj.magic (-1)
-external text_get_text : tree -> [`Text] node -> string = "caml_text_collection_get_text"
+external text_get_tc_text : tree -> [`Text] node -> string = "caml_text_collection_get_text"
external text_is_empty : tree -> [`Text ] node -> bool = "caml_text_collection_empty_text"
let text_is_empty t n =
(equal_node nil n) || text_is_empty t n
-external get_cached_text : tree -> [`Text ] node -> string = "caml_text_collection_get_cached_text"
-
-
-let text_get_text t n =
- if equal_node nil n then ""
- else get_cached_text t n
-external text_size : tree -> int = "caml_text_collection_size"
+
external text_is_contains : tree -> string -> bool = "caml_text_collection_is_contains"
external text_count_contains : tree -> string -> int = "caml_text_collection_count_contains"
external text_count : tree -> string -> int = "caml_text_collection_count"
external text_contains : tree -> string -> [`Text ] node array = "caml_text_collection_contains"
+external text_unsorted_contains : tree -> string -> unit = "caml_text_collection_unsorted_contains"
+external text_get_cached_text : tree -> [`Text] node -> string = "caml_text_collection_get_cached_text"
-external tree_serialize : tree -> string -> unit = "caml_xml_tree_serialize"
+external tree_serialize : tree -> string -> unit = "caml_xml_tree_serialize"
external tree_unserialize : string -> tree = "caml_xml_tree_unserialize"
external tree_parent_doc : tree -> [`Text ] node -> [`Tree ] node = "caml_xml_tree_parent_doc"
external tree_prev_doc : tree -> [`Text ] node -> [`Tree ] node = "caml_xml_tree_prev_doc"
external tree_first_child : tree -> [`Tree] node -> [`Tree] node = "caml_xml_tree_first_child"
+external tree_tagged_child : tree -> [`Tree] node -> Tag.t -> [`Tree] node = "caml_xml_tree_tagged_child"
external tree_next_sibling : tree -> [`Tree] node -> [`Tree] node = "caml_xml_tree_next_sibling"
+external tree_tagged_sibling : tree -> [`Tree] node -> Tag.t -> [`Tree] node = "caml_xml_tree_tagged_sibling"
+
external tree_prev_sibling : tree -> [`Tree] node -> [`Tree] node = "caml_xml_tree_prev_sibling"
external tree_is_leaf : tree -> [`Tree] node -> bool = "caml_xml_tree_is_leaf"
-
+external tree_last_child : tree -> [`Tree] node -> [`Tree] node = "caml_xml_tree_last_child"
+external tree_is_first_child : tree -> [`Tree] node -> bool = "caml_xml_tree_is_first_child"
+
(* external tag : tree -> [`Tree ] node -> T = "caml_xml_tree_tag"*)
external tree_tag_id : tree -> [`Tree ] node -> Tag.t = "caml_xml_tree_tag_id"
+
let tree_is_last t n = equal_node nil (tree_next_sibling t n)
external tree_my_text : tree -> [`Tree] node -> [`Text ] node = "caml_xml_tree_my_text"
external tree_next_text : tree -> [`Tree] node -> [`Text ] node = "caml_xml_tree_next_text"
external tree_doc_ids : tree -> [`Tree ] node -> [`Text ] node * [`Text ] node = "caml_xml_tree_doc_ids"
+
+let text_size tree = int_of_node (snd ( tree_doc_ids tree (Obj.magic 0) ))
+
+let text_get_cached_text t x =
+ if x == -1 then ""
+ else
+ text_get_cached_text t x
+
+
external tree_text_xml_id : tree -> [`Text ] node -> int = "caml_xml_tree_text_xml_id"
external tree_node_xml_id : tree -> [`Tree ] node -> int = "caml_xml_tree_node_xml_id"
external tree_is_ancestor : tree -> [`Tree ] node -> [`Tree ] node -> bool = "caml_xml_tree_is_ancestor"
external tree_tagged_desc : tree -> [`Tree ] node -> Tag.t -> [`Tree ] node = "caml_xml_tree_tagged_desc"
external tree_tagged_foll_below : tree -> [`Tree ] node -> Tag.t -> [`Tree ] node -> [`Tree ] node = "caml_xml_tree_tagged_foll_below"
external tree_subtree_tags : tree -> [`Tree ] node -> Tag.t -> int = "caml_xml_tree_subtree_tags"
-external tree_select_below : tree -> [`Tree ] node -> Ptset.int_vector -> Ptset.int_vector -> [`Tree ] node = "caml_xml_tree_select_below"
-external tree_select_desc_only : tree -> [`Tree ] node -> Ptset.int_vector -> [`Tree ] node = "caml_xml_tree_select_desc_only"
-external tree_select_next : tree -> [`Tree ] node -> Ptset.int_vector -> Ptset.int_vector -> [`Tree ] node -> [`Tree ] node = "caml_xml_tree_select_next"
-external tree_select_foll_only : tree -> [`Tree ] node -> Ptset.int_vector -> [`Tree ] node -> [`Tree ] node = "caml_xml_tree_select_foll_only"
-external tree_select_desc_or_foll_only : tree -> [`Tree ] node -> Ptset.int_vector -> [`Tree ] node -> [`Tree ] node = "caml_xml_tree_select_foll_only"
-
-type descr =
- | Nil
- | Node of [`Tree] node
- | Text of [`Text] node * [`Tree] node
+
+
+
+type int_vector
+external int_vector_alloc : int -> int_vector = "caml_int_vector_alloc"
+external int_vector_length : int_vector -> int = "caml_int_vector_length"
+external int_vector_set : int_vector -> int -> int -> unit = "caml_int_vector_set"
+
+external tree_select_child : tree -> [`Tree ] node -> int_vector -> [`Tree] node = "caml_xml_tree_select_child"
+external tree_select_foll_sibling : tree -> [`Tree ] node -> int_vector -> [`Tree] node = "caml_xml_tree_select_foll_sibling"
+external tree_select_desc : tree -> [`Tree ] node -> int_vector -> [`Tree] node = "caml_xml_tree_select_desc"
+external tree_select_foll_below : tree -> [`Tree ] node -> int_vector -> [`Tree] node -> [`Tree] node = "caml_xml_tree_select_foll_below"
+
+
+module HPtset = Hashtbl.Make(Ptset.Int)
+
+let vector_htbl = HPtset.create MED_H_SIZE
+
+let ptset_to_vector s =
+ try
+ HPtset.find vector_htbl s
+ with
+ Not_found ->
+ let v = int_vector_alloc (Ptset.Int.cardinal s) in
+ let _ = Ptset.Int.fold (fun e i -> int_vector_set v i e;i+1) s 0 in
+ HPtset.add vector_htbl s v; v
+
type t = { doc : tree;
- node : descr;
- ttable : (Tag.t,(Ptset.t*Ptset.t)) Hashtbl.t;
+ node : [`Tree] node;
+ ttable : (Tag.t,(Ptset.Int.t*Ptset.Int.t)) Hashtbl.t;
}
-
-let update h t sb sa =
+let text_size t = text_size t.doc
+
+module MemUnion = Hashtbl.Make (struct
+ type t = Ptset.Int.t*Ptset.Int.t
+ let equal (x,y) (z,t) = (Ptset.Int.equal x z)&&(Ptset.Int.equal y t)
+ let equal a b = equal a b || equal b a
+ let hash (x,y) = (* commutative hash *)
+ let x = Ptset.Int.hash x
+ and y = Ptset.Int.hash y
+ in
+ if x < y then HASHINT2(x,y) else HASHINT2(y,x)
+ end)
+
+let collect_tags tree =
+ let h_union = MemUnion.create BIG_H_SIZE in
+ let pt_cup s1 s2 =
+ try
+ MemUnion.find h_union (s1,s2)
+ with
+ | Not_found -> let s = Ptset.Int.union s1 s2
+ in
+ MemUnion.add h_union (s1,s2) s;s
+ in
+ let h_add = Hashtbl.create BIG_H_SIZE in
+ let pt_add t s =
+ let k = HASHINT2(Tag.hash t,Ptset.Int.hash s) in
+ try
+ Hashtbl.find h_add k
+ with
+ | Not_found -> let r = Ptset.Int.add t s in
+ Hashtbl.add h_add k r;r
+ in
+ let h = Hashtbl.create BIG_H_SIZE in
+ let update t sb sa =
let sbelow,safter =
try
Hashtbl.find h t
with
- | Not_found -> Ptset.empty,Ptset.empty
+ | Not_found ->
+ (Ptset.Int.empty,Ptset.Int.empty)
in
- Hashtbl.replace h t (Ptset.union sbelow sb, Ptset.union safter sa)
-
-
-
-
-
-let collect_tags tree =
- let h = Hashtbl.create 511 in
+ Hashtbl.replace h t (pt_cup sbelow sb, pt_cup safter sa)
+ in
let rec loop id acc =
if equal_node id nil
- then (Ptset.singleton Tag.pcdata, Ptset.add Tag.pcdata acc)
+ then (Ptset.Int.empty,acc)
else
let below2,after2 = loop (tree_next_sibling tree id) acc in
let below1,after1 = loop (tree_first_child tree id) after2 in
let tag = tree_tag_id tree id in
- update h tag below1 after2;
- Ptset.add tag (Ptset.union below1 below2), (Ptset.add tag after1)
+ update tag below1 after2;
+ pt_add tag (pt_cup below1 below2), (pt_add tag after1)
in
- let b,a = loop (tree_root tree) Ptset.empty in
- update h Tag.pcdata b a;
- h
+ let _ = loop (tree_root tree) Ptset.Int.empty in h
+
+
let contains_array = ref [| |]
-
+let contains_index = Hashtbl.create 4096
+let in_array _ i =
+ try
+ Hashtbl.find contains_index i
+ with
+ Not_found -> false
+
let init_contains t s =
let a = text_contains t.doc s
in
Array.fast_sort (compare) a;
- contains_array := a
+ contains_array := a;
+ Array.iter (fun x -> Hashtbl.add contains_index x true) !contains_array
+let count_contains t s = text_count_contains t.doc s
+let unsorted_contains t s = text_unsorted_contains t.doc s
+
let init_naive_contains t s =
let i,j = tree_doc_ids t.doc (tree_root t.doc)
in
let rec loop n acc l =
if n >= j then acc,l
else
- let s = text_get_text t.doc n
+ let s = text_get_cached_text t.doc n
in
if matching s
then loop (n+1) (n::acc) (l+1)
let compare = compare_node end)
end
-let is_nil t = match t.node with
- | Nil -> true
- | Node(i) -> equal_node i nil
- | _ -> false
-
-let is_node t =
-match t.node with
- | Node(i) -> not(equal_node i nil)
- | _ -> false
+let is_nil t = t.node == nil
+let is_node t = t.node != nil
-let node_of_t t =
+let node_of_t t =
let _ = Tag.init (Obj.magic t) in
let table = collect_tags t
in
-(*
- let _ = Hashtbl.iter (fun t (sb,sa) ->
- Printf.eprintf "'%s' -> { " (Tag.to_string t);
- Ptset.iter (fun i -> Printf.eprintf "'%s' " (Tag.to_string i)) sb;
- Printf.eprintf "}\n { ";
- Ptset.iter (fun i -> Printf.eprintf "'%s' " (Tag.to_string i)) sa;
- Printf.eprintf "} \n----------------------------------\n";
- ) table in
-*)
{ doc= t;
- node = Node(tree_root t);
+ node = tree_root t;
ttable = table;
}
+let finalize _ = Printf.eprintf "Release the string list !\n%!"
+;;
+
+let parse f str =
+ node_of_t
+ (f str
+ !Options.sample_factor
+ !Options.index_empty_texts
+ !Options.disable_text_collection)
+
+let parse_xml_uri str = parse parse_xml_uri str
+let parse_xml_string str = parse parse_xml_string str
-
-let parse_xml_uri str = node_of_t
- (parse_xml_uri str
- !Options.sample_factor
- !Options.index_empty_texts
- !Options.disable_text_collection)
-
-let parse_xml_string str = node_of_t
- (parse_xml_string str
- !Options.sample_factor
- !Options.index_empty_texts
- !Options.disable_text_collection)
external pool : tree -> Tag.pool = "%identity"
-let save t str = save_tree t.doc str
+
+let save t str = (save_tree t.doc str)
+;;
+
let load ?(sample=64) str =
node_of_t (load_tree str sample)
let tag_pool t = pool t.doc
-let compare a b = match a.node,b.node with
- | Nil, Nil -> 0
- | Nil,_ -> 1
- | _ , Nil -> -1
- | Node(i),Node(j) -> compare_node i j
- | Text(i,_), Text(j,_) -> compare_node i j
- | Node(i), Text(_,j) -> compare_node i j
- | Text(_,i), Node(j) -> compare_node i j
-
-let equal a b = (compare a b) == 0
-
-
-let norm (n : [`Tree ] node ) = if tree_is_nil n then Nil else Node (n)
-
+let compare a b = a.node - b.node
+
+let equal a b = a.node == b.node
+
let nts = function
- Nil -> "Nil"
- | Text (i,j) -> Printf.sprintf "Text (%i, %i)" i j
- | Node (i) -> Printf.sprintf "Node (%i)" i
+ -1 -> "Nil"
+ | i -> Printf.sprintf "Node (%i)" i
-let mk_nil t = { t with node = Nil }
-let root n = { n with node = norm (tree_root n.doc) }
+let dump_node t = nts t.node
-let is_root n = match n.node with
- | Node(t) -> (int_of_node t) == 0
- | _ -> false
-
-let parent n =
- let node' =
- match n.node with (* inlined parent *)
- | Node(t) when (int_of_node t)== 0 -> Nil
- | Node(t) ->
- let txt = tree_prev_text n.doc t in
- if text_is_empty n.doc txt then
- let ps = tree_prev_sibling n.doc t in
- if tree_is_nil ps
- then
- Node(tree_parent n.doc t)
- else Node(ps)
- else
- Text(txt,t)
- | Text(i,t) ->
- let ps = tree_prev_doc n.doc i in
- if tree_is_nil ps
- then Node (tree_parent_doc n.doc i)
- else Node(ps)
- | _ -> failwith "parent"
- in
- { n with node = node' }
-
-let node_child n =
- match n.node with
- | Node i -> { n with node= norm(tree_first_child n.doc i) }
- | _ -> { n with node = Nil }
-
-let node_sibling n =
- match n.node with
- | Node i -> { n with node= norm(tree_next_sibling n.doc i) }
- | _ -> { n with node = Nil }
-
-let node_sibling_ctx n _ =
- match n.node with
- | Node i -> { n with node= norm(tree_next_sibling n.doc i) }
- | _ -> { n with node = Nil }
-
-
-let first_child n =
- let node' =
- match n.node with
- | Node (t) ->
- let fs = tree_first_child n.doc t in
- if equal_node nil fs
- then
- let txt = tree_my_text n.doc t in
- if equal_node nil txt
- then Nil
- else Text(txt,nil)
- else
- let txt = tree_prev_text n.doc fs in
- if equal_node nil txt
- then Node(fs)
- else Text(txt, fs)
- | Text(_,_) -> Nil
- | Nil -> failwith "first_child"
- in
- { n with node = node'}
+let mk_nil t = { t with node = nil }
+let root n = { n with node = tree_root n.doc }
+
+let is_root n = n.node == (tree_root n.doc)
-let next_sibling n =
- let node' =
- match n.node with
- | Text (_,ns) -> norm ns
- | Node(t) ->
- let ns = tree_next_sibling n.doc t in
- let txt = tree_next_text n.doc t in
- if equal_node nil txt
- then norm ns
- else Text(txt, ns)
- | Nil -> failwith "next_sibling"
- in
- { n with node = node'}
-
-let next_sibling_ctx n _ = next_sibling n
-
-let left = first_child
-let right = next_sibling
-
-let id t =
- match t.node with
- | Node(n) -> tree_node_xml_id t.doc n
- | Text(i,_) -> tree_text_xml_id t.doc i
- | _ -> -1
-
-let tag t =
- match t.node with
- | Text(_) -> Tag.pcdata
- | Node(n) -> tree_tag_id t.doc n
- | _ -> failwith "tag"
-
-(*
- let string_below t id =
- let strid = parent_doc t.doc id in
- match t.node with
- | Node(NC(i)) ->
- (Tree.equal i strid) || (is_ancestor t.doc i strid)
- | Node(SC(i,_)) -> Text.equal i id
- | _ -> false
-
-
- let tagged_foll t tag =
- if tag = Tag.attribute || tag = Tag.pcdata then failwith "tagged_foll"
- else match t with
- | { doc=d; node=Node(NC n) } -> { t with node = norm (tagged_foll d n tag) }
- | { doc=d; node=Node(SC (_,n)) } when is_nil n -> { t with node= Nil }
- | { doc=d; node=Node(SC (_,n)) } ->
- let nnode =
- if tag_id d n == tag then n
- else
- let n' = tagged_desc d n tag in
- if is_nil n' then tagged_foll d n tag
- else n'
- in {t with node= norm nnode}
- | _ -> { t with node=Nil }
-
+let is_left n = tree_is_first_child n.doc n.node
- let tagged_desc t tag =
- if tag = Tag.attribute || tag = Tag.pcdata then failwith "tagged_desc"
- else match t with
- | { doc=d; node=Node(NC n) } -> { t with node = norm (tagged_desc d n tag) }
- | _ -> { t with node=Nil }
-
-*)
-let select_next tb tf t s =
- match s.node with
- | Node (below) -> begin
- match t.node with
- | Node( n) ->
- { t with node = norm (tree_select_next t.doc n (Ptset.to_int_vector tb) (Ptset.to_int_vector tf) below) }
- | Text (i,n) when equal_node nil n ->
- let p = tree_parent_doc t.doc i in
- { t with node = norm (tree_select_next t.doc p (Ptset.to_int_vector tb) (Ptset.to_int_vector tf) below) }
- | Text(_,n) ->
- if Ptset.mem (tree_tag_id t.doc n) (Ptset.union tb tf)
- then { t with node=Node(n) }
- else
- let vb = Ptset.to_int_vector tb in
- let vf = Ptset.to_int_vector tf in
- let node =
- let dsc = tree_select_below t.doc n vb vf in
- if equal_node nil dsc
- then tree_select_next t.doc n vb vf below
- else dsc
- in
- { t with node = norm node }
- | _ -> {t with node = Nil }
- end
-
- | _ -> { t with node = Nil }
+let is_below_right t1 t2 = tree_is_ancestor t1.doc (tree_parent t1.doc t1.node) t2.node
-
+let parent n = { n with node = tree_parent n.doc n.node }
+let first_child n = { n with node = tree_first_child n.doc n.node }
+let tagged_child tag n = { n with node = tree_tagged_child n.doc n.node tag }
+let select_child ts n = { n with node = tree_select_child n.doc n.node (ptset_to_vector ts) }
- let select_foll_only tf t s =
- match s.node with
- | Node (below) ->
- begin
- match t.node with
- | Node(n) ->
- { t with node= norm (tree_select_foll_only t.doc n (Ptset.to_int_vector tf) below) }
- | Text(i,n) when equal_node nil n ->
- let p = tree_parent_doc t.doc i in
- { t with node= norm (tree_select_foll_only t.doc p (Ptset.to_int_vector tf) below) }
- | Text(_,n) ->
- if Ptset.mem (tree_tag_id t.doc n) tf
- then { t with node=Node(n) }
- else
- let vf = Ptset.to_int_vector tf in
- let node =
- let dsc = tree_select_desc_only t.doc n vf in
- if tree_is_nil dsc
- then tree_select_foll_only t.doc n vf below
- else dsc
- in
- { t with node = norm node }
- | _ -> { t with node = Nil }
- end
- | _ -> {t with node=Nil }
-
-let select_below tc td t=
- match t.node with
- | Node( n) ->
- let vc = Ptset.to_int_vector tc
- in
- let vd = Ptset.to_int_vector td
- in
- { t with node= norm(tree_select_below t.doc n vc vd) }
- | _ -> { t with node=Nil }
-
+let next_sibling n = { n with node = tree_next_sibling n.doc n.node }
+let tagged_sibling tag n = { n with node = tree_tagged_sibling n.doc n.node tag }
+let select_sibling ts n = { n with node = tree_select_foll_sibling n.doc n.node (ptset_to_vector ts) }
+
+let next_sibling_ctx n _ = next_sibling n
+let tagged_sibling_ctx tag n _ = tagged_sibling tag n
+let select_sibling_ctx ts n _ = select_sibling ts n
+
+let id t = tree_node_xml_id t.doc t.node
-let select_desc_only td t =
- match t.node with
- | Node(n) ->
- let vd = Ptset.to_int_vector td
- in
- { t with node = norm(tree_select_desc_only t.doc n vd) }
- | _ -> { t with node = Nil }
-
-
-let tagged_desc tag t =
- match t.node with
- | Node(n) ->
- { t with node = norm(tree_tagged_desc t.doc n tag) }
- | _ -> { t with node = Nil }
-
-
-let tagged_foll_below tag t s =
- match s.node with
- | Node (below) ->
- begin
- match t.node with
- | Node(n) ->
- { t with node= norm (tree_tagged_foll_below t.doc n tag below) }
- | Text(i,n) when equal_node nil n ->
- let p = tree_prev_doc t.doc i in
- { t with node= norm (tree_tagged_foll_below t.doc p tag below) }
- | Text(_,n) ->
- if (tree_tag_id t.doc n) == tag
- then { t with node=Node(n) }
- else
- let node =
- let dsc = tree_tagged_desc t.doc n tag in
- if tree_is_nil dsc
- then tree_tagged_foll_below t.doc n tag below
- else dsc
- in
- { t with node = norm node }
- | _ -> { t with node = Nil }
- end
- | _ -> {t with node=Nil }
+let tag t = if t.node == nil then Tag.nullt else tree_tag_id t.doc t.node
+let tagged_desc tag n = { n with node = tree_tagged_desc n.doc n.node tag }
+let select_desc ts n = { n with node = tree_select_desc n.doc n.node (ptset_to_vector ts) }
+
+let tagged_foll_ctx tag t ctx =
+ { t with node = tree_tagged_foll_below t.doc t.node tag ctx.node }
+let select_foll_ctx ts n ctx = { n with node = tree_select_foll_below n.doc n.node (ptset_to_vector ts) ctx.node }
let last_idx = ref 0
let array_find a i j =
else loop !last_idx i j
-
-let text_below t =
- let l = Array.length !contains_array in
- if l = 0 then { t with node=Nil }
- else
- match t.node with
- | Node(n) ->
- let i,j = tree_doc_ids t.doc n in
- let id = array_find !contains_array i j
- in
- if id == nil then
- { t with node=Nil }
- else
- { t with node = Text(id, tree_next_sibling t.doc (tree_prev_doc t.doc id)) }
- | _ -> { t with node = Nil }
-
-let text_next t root =
- let l = Array.length !contains_array in
- if l = 0 then { t with node=Nil }
- else
- let inf = match t.node with
- | Node(n) -> snd(tree_doc_ids t.doc n)+1
- | Text(i,_) -> i+1
- | _ -> assert false
- in
- match root.node with
- | Node (n) ->
- let _,j = tree_doc_ids t.doc n in
- let id = array_find !contains_array inf j
- in
- if id == nil then { t with node= Nil }
- else
- { t with node = Text(id,tree_next_sibling t.doc (tree_prev_doc t.doc id)) }
- | _ -> { t with node = Nil}
-
-
-(*
- let subtree_tags t tag =
- match t with
- { doc = d; node = Node(NC n) } ->
- subtree_tags d n tag
- | _ -> 0
-
- let select_desc_array = ref [| |]
- let idx = ref 0
-
- let init_tagged_next t tagid =
- let l = subtree_tags (root t) tagid
- in
- tagged_desc_array := Array.create l { t with node= Nil };
- let i = ref 0 in
- let rec collect t =
- if is_node t then begin
- if tag t == tagid then
- begin
- !tagged_desc_array.(!i) <- t;
- incr i;
- end;
- collect (first_child t);
- collect (next_sibling t)
- end;
- in
- collect t;
- idx := 0
-
- let print_id ppf v =
- let pr x= Format.fprintf ppf x in
- match v with
- { node=Nil } -> pr "NULLT: -1"
- | { node=String(i) } | { node=Node(SC(i,_)) } -> pr "DocID: %i" (int_of_node i)
- | { node=Node(NC(i)) } -> pr "Node: %i" (int_of_node i)
-
-
-
-(* let tagged_next t tag =
- if !idx >= Array.length !tagged_desc_array
- then {t with node=Nil}
- else
- let r = !tagged_desc_array.(!idx)
- in
- incr idx; r
-*)
-
-
- let has_tagged_foll t tag = is_node (tagged_foll t tag)
- let has_tagged_desc t tag = is_node (tagged_desc t tag)
-
- let contains t s =
- Array.fold_left (fun a i -> DocIdSet.add i a) DocIdSet.empty (Text.contains t.doc s)
-
-
- let contains_old t s =
- let regexp = Str.regexp_string s in
- let matching arg =
- try
- let _ = Str.search_forward regexp arg 0;
- in true
- with _ -> false
- in
- let rec find t acc = match t.node with
- | Nil -> acc
- | String i ->
- if matching (string t) then DocIdSet.add i acc else acc
- | Node(_) -> (find (left t )) ((find (right t)) acc)
- in
- find t DocIdSet.empty
-
-
- let contains_iter t s =
- let regexp = Str.regexp_string s in
- let matching arg =
- try
- let _ = Str.search_forward regexp arg 0;
- in true
- with _ -> false
- in
- let size = Text.size t.doc in
- let rec find acc n =
- if n == size then acc
- else
- find
- (if matching (Text.get_cached_text t.doc (Obj.magic n)) then
- DocIdSet.add (Obj.magic n) acc
- else acc) (n+1)
- in
- find DocIdSet.empty 0
-
-
-
-
- let count_contains t s = Text.count_contains t.doc s
-*)
let count t s = text_count t.doc s
-(*
- let is_left t =
- if is_root t then false
- else
- if tag (parent t) == Tag.pcdata then false
- else
- let u = left (parent t) in
- (id t) == (id u)
-*)
+
let print_xml_fast outc t =
let rec loop ?(print_right=true) t =
- match t.node with
- | Nil -> ()
- | Text(i,n) -> output_string outc (text_get_text t.doc i);
+ if t.node != nil
+ then
+ let tagid = tree_tag_id t.doc t.node in
+ if tagid==Tag.pcdata
+ then output_string outc (text_get_cached_text t.doc t.node);
if print_right
- then loop (right t)
- | Node (n) ->
- let tg = Tag.to_string (tag t) in
- let l = left t
- and r = right t
- in
- output_char outc '<';
- output_string outc tg;
- ( match l.node with
- Nil -> output_string outc "/>"
- | Node(_) when Tag.equal (tag l) Tag.attribute ->
- (loop_attributes (left l);
- match (right l).node with
- | Nil -> output_string outc "/>"
- | _ ->
- output_char outc '>';
- loop (right l);
- output_string outc "</";
- output_string outc tg;
- output_char outc '>' )
- | _ ->
+ then loop (next_sibling t)
+
+ else
+ let tagstr = Tag.to_string tagid in
+ let l = first_child t
+ and r = next_sibling t
+ in
+ output_char outc '<';
+ output_string outc tagstr;
+ if l.node == nil then output_string outc "/>"
+ else
+ if (tag l) == Tag.attribute then
+ begin
+ loop_attributes (first_child l);
+ if (next_sibling l).node == nil then output_string outc "/>"
+ else
+ begin
+ output_char outc '>';
+ loop (next_sibling l);
+ output_string outc "</";
+ output_string outc tagstr;
+ output_char outc '>';
+ end;
+ end
+ else
+ begin
output_char outc '>';
loop l;
output_string outc "</";
- output_string outc tg;
- output_char outc '>'
- );if print_right then loop r
- and loop_attributes a =
- match a.node with
- | Node(_) ->
- let value =
- match (left a).node with
- | Text(i,_) -> text_get_text a.doc i
- | _ -> assert false
- in
- output_char outc ' ';
- output_string outc (Tag.to_string (tag a));
- output_string outc "=\"";
- output_string outc value;
- output_char outc '"';
- loop_attributes (right a)
- | _ -> ()
+ output_string outc tagstr;
+ output_char outc '>';
+ end;
+ if print_right then loop r
+ and loop_attributes a =
+ let s = (Tag.to_string (tag a)) in
+ let attname = String.sub s 3 ((String.length s) -3) in
+ output_char outc ' ';
+ output_string outc attname;
+ output_string outc "=\"";
+ output_string outc (text_get_cached_text t.doc
+ (tree_my_text a.doc (first_child a).node));
+ output_char outc '"';
+ loop_attributes (next_sibling a)
in
loop ~print_right:false t
let print_xml_fast outc t =
- if Tag.to_string (tag t) = "" then
+ if (tag t) = Tag.document_node then
print_xml_fast outc (first_child t)
- else print_xml_fast outc t
+ else print_xml_fast outc t
let tags_below t tag =
snd(Hashtbl.find t.ttable tag)
let tags t tag = Hashtbl.find t.ttable tag
+
+
+let rec binary_parent t =
+ if tree_is_first_child t.doc t.node
+ then { t with node = tree_parent t.doc t.node }
+ else { t with node = tree_prev_sibling t.doc t.node }
+
+let doc_ids (t:t) : (int*int) =
+ (Obj.magic (tree_doc_ids t.doc t.node))
+
+let subtree_tags t tag =
+ if t.node == nil then 0 else
+ tree_subtree_tags t.doc t.node tag
+
+let get_text t =
+ let tid = tree_my_text t.doc t.node in
+ if tid == nil then "" else
+ let a, b = tree_doc_ids t.doc (tree_root t.doc) in
+ let _ = Printf.eprintf "Trying to take text %i of node %i in %i %i\n%!" tid t.node a b in
+ text_get_cached_text t.doc tid
+
+
+let dump_tree fmt t =
+ let rec loop tree n =
+ if tree != nil then
+ let tag = (tree_tag_id t.doc tree ) in
+ let tagstr = Tag.to_string tag in
+ let tab = String.make n ' ' in
+
+ if tag == Tag.pcdata || tag == Tag.attribute_data
+ then
+ Format.fprintf fmt "%s<%s>%s</%s>\n"
+ tab tagstr (text_get_cached_text t.doc (tree_my_text t.doc tree)) tagstr
+ else begin
+ Format.fprintf fmt "%s<%s>\n" tab tagstr;
+ loop (tree_first_child t.doc tree) (n+2);
+ Format.fprintf fmt "%s</%s>\n%!" tab tagstr;
+ end;
+ loop (tree_next_sibling t.doc tree) n
+ in
+ loop (tree_root t.doc) 0
+;;
+
+